High-Resolution SNP/CGH Microarrays Reveal the Accumulation of Loss of Heterozygosity in Commonly Used Candida albicans Strains

Abstract

Phenotypic diversity can arise rapidly through loss of heterozygosity (LOH) or by the acquisition of copy number variations (CNV) spanning whole chromosomes or shorter contiguous chromosome segments. In Candida albicans, a heterozygous diploid yeast pathogen with no known meiotic cycle, homozygosis and aneuploidy alter clinical characteristics, including drug resistance. Here, we developed a high-resolution microarray that simultaneously detects ∼39,000 single nucleotide polymorphism (SNP) alleles and ∼20,000 copy number variation loci across the C. albicans genome. An important feature of the array analysis is a computational pipeline that determines SNP allele ratios based upon chromosome copy number. Using the array and analysis tools, we constructed a haplotype map (hapmap) of strain SC5314 to assign SNP alleles to specific homologs, and we used it to follow the acquisition of loss of heterozygosity (LOH) and copy number changes in a series of derived laboratory strains. This high-resolution SNP/CGH microarray and the associated hapmap facilitated the phasing of alleles in lab strains and revealed detrimental genome changes that arose frequently during molecular manipulations of laboratory strains. Furthermore, it provided a useful tool for rapid, high-resolution, and cost-effective characterization of changes in allele diversity as well as changes in chromosome copy number in new C. albicans isolates.

Keywords: aneuploidy; comparative genome hybridization; genome profiling; haplotype mapping; loss of heterozygosity; single nucleotide polymorphisms.

Figure 1

Homolog genotypes of strains used for hapmap construction. Strains derived from a parasexual cross (Forche et al. 2008) contained homozygous (“aa” or “bb”) homologs for most or all of ChrR and Chr2–Chr7 and heterozygous trisomies (“aab” or “abb”) as indicated. Chromosomes indicated with “aab“ or “abb“ were trisomic; chromosomes designated “ab” were heterozygous along the entire length of the chromosome. Chromosomes indicated in color were informative for hapmap construction. Chr2 in strain P1 (**) was homozygous for each homolog along different sections of the chromosome. Color schemes for allelic fractions are consistent in Figures 1–4.

Figure 2

Assignment of SNPs from allele hybridization ratios. Allelic fractions were calculated from SNP array hybridization data for each chromosome in each strain used to construct the hapmap. In each case, unassigned alleles (left panels with only gray data) were plotted with the X-axis as the allelic fraction [b/(a+b)] and the Y-axis as the number of SNP probes, scaled to maximum height of the peaks. (A) For unassigned alleles on heterozygous chromosomes, a = b and b/(a+b) = 0.5 (dark black vertical line); probes that fall between the boundaries (right panel, green lines, values determined as described in Material and Methods) are considered heterozygous (gray); probes outside these loci are considered homozygous. (B) For unassigned alleles on homozygous chromosomes, the allelic fractions were close to either 0 or 1 (left panel). For alleles on chromosomes previously designated “aa,” SNP pairs with fractions that fell in the “bb” peak were reassigned to “aa” (magenta, middle panel), and vice versa for “bb” alleles (cyan, right panel) using boundaries that were 50% of the distance between the two peaks in the unassigned data and the heterozygous peak location (green lines). (C) For unassigned alleles on trisomic heterozygous chromosomes, the allelic fractions were close to ∼0.33 or ∼0.67 (“aab” or “abb,” respectively, left panel). SNP pairs with fractions that fell in the “abb” peak were reassigned to “aab” (purple, middle panel), and vice-versa for “aab” alleles (blue, right panel). The boundary between the two regions was set at 50% of the distance between the two peaks (red lines). SNP, single nucleotide polymorphism.

Figure 3

Visualization of SNP/CGH array data. Data for each strain analyzed by SNP/CGH array was visualized as illustrated for strain Ps8 (and in Figure S1 and Figure S2 for all other strains). Each chromosome is illustrated to scale, with its centromere position indicated by an indentation and major repeat sequence (MRS) positions indicted by a black square below the chromosome. CGH data were calculated as ratios, converted to chromosome copy numbers, displayed as black histograms along the length of the chromosome, and summarized with a large numeral to the right of the chromosome. For SNP alleles, allelic fractions for each chromosome are plotted in the left panels. SNP data were calculated and colored as described in Figure 2. Regions that were homozygous in the reference strain or were not informative were not colored. CGH, comparative genome hybridization; SNP, single nucleotide polymorphism.

Figure 4

Genome changes acquired during laboratory strain construction. Summary of SNP/CGH analyses of laboratory strains derived from SC5314 with chromosome copy number illustrated as black histograms and homolog identity colored as in Figure 3. Each strain is illustrated horizontally. Strain construction steps proceed from top to bottom. RM100#13 (bottom of A) was the parent for strain lineages in B and C. Genome changes with successive transformation and counterselection steps are enlarged and/or highlighted: new aneuploidies are highlighted with green borders; new LOH events are highlighted with yellow borders. Genome changes that were maintained in subsequent strains are highlighted with black borders. Subfigures for strains with italicized names were simulated from available data and SNP/CGH analysis of additional CAI-4-derived strains. CGH, comparative genome hybridization; LOH, loss of heterozygosity; SNP, single nucleotide polymorphism.

Figure 5

In vitro growth of lab strains with different degrees of ploidy or LOH or with different URA3 copy number. (A) Deviations from diploidy correlate with increased doubling times. Doubling times (min) were measured for strains with similar degrees of LOH and different degrees whole-genome ploidy. Blue: CAF2 (2N), CAI-4 F2 [2.16N (2N+Chr2)] and CAI-4 F3 [2.38 (2N + (Chr1 + Chr2))] Green: RM1000 #2 (2N) and RM1000 #6 (1.997N) have identical genome content but are homozygous or hemizygous, respectively, for the terminal ∼40 kb of Chr5R. Red: SN76 (2N) and BWP17 (1.997N), like the RM1000 strains, are homozygous and hemizygous for the tip of Chr5R; in addition, BWP17 acquired a small (30 kb) LOH on Chr4R. *Within each group, all strains are significantly different from each other (P < 0.05, Student t-test). (B) Increases in percentage of the genome that is homozygous (as indicated on X-axis) correlate with increased doubling times. SC5314 and CAF2 are not different from each other (+) but are significantly different from RM100 #13 and SN76 (*P < 0.05, Student t-test). (C) URA3 status does not affect in vitro doubling times in medium containing uridine. Pairs of strains (teal and purple) with similar ploidies and similar degrees of LOH that differ by one copy of URA3: SC5314 (URA3/URA3) vs. CAF2 (URA3/ura3∆), or SN76 (ura3∆/ura3∆) vs. SN95 (URA3/ura3)) were not significant different in doubling time relative to one another (+). LOH, loss ofheterozygosity.